shuttie commented on issue #10529: [FLINK-15171] [serialization] fix 
performance regression caused by too many buffer allocations on string 
serialization
URL: https://github.com/apache/flink/pull/10529#issuecomment-564530295
 
 
   A good follow-up question is "how large should be the 
`SHORT_STRING_MAX_LENGTH` constant"? Currently it's 1024, as I guess having a 
per-thread overhead of +3kb is indistinguishable from noise, but gives a nice 
performance boost.
   
   It can be trivially bumped to something like 64kb without any drawbacks, but 
from my personal experience with Apache Flink, giant strings are quite a rare 
case in real world projects. And for 1kb+ strings, the original PR already 
improves the situation.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to